|NDT.net Oct 2005 Vol. 10 No.10|
Fibre Reinforced Plastic: A Feasibility Study of Microwave Based Non-Destructive TestingD. Beilken
FI Test- und Messtechnik GmbH, Magdeburg, Germany
J. H. Hinken
AbstractIn the non-destructive testing (NDT) of glass- and natural-fibre reinforced plastic composites special problems may arise due to their inhomogeneous nature. Standard NDT methods may fail or have important limitations. This feasibility study shows that microwave based NDT lends itself as a powerful tool for this task. Four example measurements on samples with artifical defects are described: profiles of endless-fibre reinforced plastics, a PTFE plate as well as natural- and glass-fibre reinforced plastic plates.
1. IntroductionPresently, many constructions are made from composite materials instead of metal in order to save weight. For example rotor blades of wind-powered devices or of helicopters are made of composite glass fibre materials. A similar situation is the spray on foam isolation at a space shuttle . In order to ensure the safety of these objects, it is necessary to be able to test the material parameters ranging from the production of the raw material up to the finished structure and also after possible damages caused by ageing or accident. Principally only few non-destructive techniques can be used to test composite glass fibre structures. These are modified ultrasonic or thermographic methods, both having their limitations.
The problems of finding defects in composite materials is caused by inhomogeneity of the structure: composite material is plastic material with innumerable inclusions in filament form and boundary surfaces. Genuine defects such as foreign material inclusions, unbonds , voids and tears among other types of defects in the inhomogeneous structure have to be recognised. Obviously, more different types of defects can arise than within homogeneous materials. Only in the rarest cases the defects are visible with the naked eye. Impact defects furthermore, tend to be wider spread in the depth of the device than at its surface. That makes it necessary to use non-destructive testing techniques for inspection.
In  it was shown that microwave and millimeter wave based non-destructive testing has a great potential for the inspection of fibre reinforced plastics. The samples are irradiated by weak microwaves, and the reflected signals are measured. This contains information on the spatial distribution of the dielectric constant of the sample and, therefore, on the internal condition of it.
2. Measurement setupIn this feasibility study our scanning system consisted of a plotter used as an x-y-stage that supports and moves the sample under a fixed open ended waveguide used as the transmitting and receiving antenna. A computer is used to generate C-scan plots from the measured data and from the x- and y-information.
The measurement setup, see fig. 1, can be splitted into two parts: a microwave part and a PC part. Basically, the latter is a PC based eddy current test system (PC4-card with ScanAlyser software from Rohmann). Another analogue/digital input/output card is used for moving the sample. The other card is used for the eddy current system. In the microwave part of the setup a tuneable 10 GHz dielectric resonator oscillator is used to generate the signal penetrating the sample. When passing the sample under the waveguide, a slowly varying signal is generated by a special kind of microwave receiver: a part of the 10 GHz generator signal via a power divider is directed to a microwave mixer. The other part is coupled through a circulator into the waveguide and its open end. By the microwave penetrating the sample placed in front of the open end, reflections are caused by spatial variations of the dielectric constant. Separated by the circulator, this reflected signal is also fed to the mixer. This mixer output signal is amplified and amplitude modulated and then fed into the PC eddy current system.
In the measurements described below, the open end of the X-band waveguide (22,86 mm x 10,16 mm) was replaced by a metallic short with a central hole (iris). This replacement increases the spatial resolution. Fig 2. shows some examples of such irises. The used one is number II in the upper right corner.
3. Measurement resultsA series of different types of samples was tested by microwave NDT. A part of the results are given below.
3.1 Central strength elements of fibre optic cablesA first investigation was performed on profiles made from endless-fibre reinforced plastic materials . Fig. 3 shows two such profiles of about 5 mm diameter axially arranged with a gap of about 1 mm in between. Two slots of 1mm depth are inserted as artificial defects into the left profile. In this single example the measurements were performed with a conducting plate behind the profiles. In the other examples the samples were supported by a holder which resembles nearly a free space situation.
The C-Scan in figure 4 shows that all three artificial defects can be recognized, the vertical V-notch in this plot only weakly, however.
3.2 PTFE plate with flat bottom holesA second example is a PTFE plate of 8 mm thickness with flat bottom holes (FBH), manufactured by the Wehrwissenschaftliches Institut für Werk-, Explosions- und Betriebstoffe (WIWEB). Fig. 5a shows a photograph of the plate taken from the side with the openings. The dimensions of the FBHs are given in Fig 5b (Durchmesser : diameter, Restwandstärke : residual wall thickness).
Fig. 5c) is a C-Scan of our microwave investigation. It shows that even small and deep lying defects can be seen.
3.3 Natural fibre reinforced plastic plateA natural fibre reinforced plastic plate made by the project group Naturstoffinnovation (ProNinA) of the University of Applied Science Magdeburg-Stendal was prepared with a row of holes with diameters of 10mm / 8mm / 6mm / 4mm / 3mm / 2mm and another row of flat bottom holes at 3mm depth and same diameters, see figure 6.
This plate was also scanned by Microwave non-destructive testing. Fig. 7 shows that the holes of dimensions down to 3mm and also the flat bottom holes in the lower row are clearly to be recognised. The holes with diameter of 2mm do not show sufficient contrast with presently used iris # II, see upper right corner in fig. 2. In some cases like overriding of the dynamic range there are artefacts surrounding the defects. But the defects themselves are represented in the scans according to their real dimensions.
3.4 Glass-fibre reinforced plastic plateA glass fibre reinforced plastic plate was produced in the Experimentelle Fabrik in Magdeburg with 1.) oil for the simulation of voids, 2.) pieces of copper and 3.) an accumulation of nature fibres, see fig. 8. This plate was scanned by Microwave NDT. Fig. 9 shows that all defects can clearly be recognised. The voids as simulated by the oil are situated on the left side of the plate. In the upper right part the accumulation of natural fibres can be seen. In the middle and diagonally to the upper right corner the buried copper strips are located.
4. ConclusionsAll scans where made using the iris # II shown in the upper right corner in fig.2. The optimisation of the iris is still under investigation. The goal is to find the best spatial resolution at a given frequency and to reduce artefacts surrounding the defects. When necessary the spatial resolution can also be improved by using higher frequencies which will make the test instruments more expensive.
To sum up, we have shown that microwave NDT is of advantage to inspect glass-fibre and natural-fibre reinforced plastic structures. Speaking more general, it is advantageous for the inspection of electrically insulating structures. Certainly, for special applications the solution of special problems may be necessary, i.e. improvement of spatial resolution, inspection speed and inspection depth. These are research goals for the future.
The authors would like to thank D. Geiss from the Polystal Composites GmbH, Th. Krell from the Wehrwissenschaftliches Institut für Werk-, Explosions- und Betriebsstoffe, P. Gerth from the Hochschule Magdeburg-Stendal (FH) and F. Laugwitz from the Experimentelle Fabrik Magdeburg for providing samples and, furthermore Th. Krell for providing measurement results. The authors also acknowledge technical support given by M. Wenk and H. Wrobel.